方法论

四项规则

Every figure on this site carries a value, a reference period, a named provider and an observation time. The reference period is the window the figure describes, not the moment we fetched it: “trailing 30 days”, “24 hours”, “annualized from the trailing 30 days”, “point-in-time”. Observation time is when the provider's value was read. Those are different things and conflating them is the most common way a data site becomes quietly wrong.

年化

Where a figure is annualized, it is scaled from the trailing 30 days by 365/30 and labeled as such. It is a run-rate, not a forecast: it states what a year would look like if the last month repeated, which it will not. We use the 30-day window rather than 24 hours because a single day of unusual activity would otherwise dominate the result, and rather than 12 months because a year-old month describes a different market.

链基础与应用基础

Fee and revenue figures are measured on one of two bases and the two are not interchangeable. A chain-basis figure is what everyone using a blockchain paid to use it. An application-basis figure is what one protocol earned from the people who used that product, summed across every chain it runs on. Each asset profile states its basis on the Fundamentals tab, and the two are never added together or ranked against each other in the same table.

Where an asset has both — a chain with a flagship application, or an application token that also secures a chain — we show the basis with the larger tracked footprint and name it, rather than combining them into a figure that describes neither.

费用、收入及两者之间的分配

Fees are the total paid by users. Revenue is the part retained by the protocol. Supply-side revenue is the remainder, paid to whoever supplied the service — miners, validators, liquidity providers, lenders. Holder revenue is the part a protocol directs to token holders by burn, buyback or distribution; it is a policy choice made by governance, not an entitlement, and it can be changed by a vote. The take rate is revenue as a share of fees.

These definitions come from the adapters our fundamentals provider maintains, and adapters differ in judgment. Two protocols that appear to do the same thing can be measured differently, which is a real limit on cross-protocol comparison and is stated on every ratio that uses these inputs.

衍生统计数据

Returns beyond one year, volatility, drawdown, beta and correlation are computed here from the daily price series we store, not taken from a provider. That way the figure and the chart a reader can look at always agree. Volatility is the standard deviation of daily log returns over the stated window, annualized by the square root of 365 — digital assets trade every day, so no trading-day convention is applied. Beta and correlation are measured against bitcoin's daily log returns over the trailing year and require at least sixty overlapping observations.

A horizon is only reported when our stored history genuinely spans it, within a fortnight's tolerance. A three-year return is left blank for an asset with two years of history rather than being approximated from a shorter window.

估值比率

Every ratio is computed from two figures already published on this site, so a reader can reconstruct any of them by hand. A ratio is only written when both legs exist and the denominator is materially non-zero; below that threshold the cell explains that the denominator is too small to divide by rather than showing a spectacular number produced by a rounding artefact.

None of these ratios means what the equity ratio of the same name means. There is no consolidated entity, no audited accounts, no residual claim and no enforceable right to any cash flow. Each definition states its own version of this and the specific way it fails.

供应量数据

Circulating supply is the weakest number on this site and it sits underneath market capitalization, dominance and every ratio built on them. It is frequently self-reported by projects and defined differently by different providers. We publish it because the alternative is publishing nothing, and we say this on every page that depends on it. Total and maximum supply are protocol facts and are far more reliable.

链上数据

Network statistics come from public explorers and node operators. Coverage is uneven by design: some chains publish a great deal without a license and some publish almost nothing. We do not fill the gaps with estimates, models or figures scraped from another chain's data. An asset whose network activity we cannot measure has a Network tab that says which figures are unavailable and why.

我们有意不发布的内容

8 figures a reader might reasonably expect are absent from this site on purpose. In each case a plausible-looking number could be produced, and in each case it would be wrong in a way the reader could not detect.

刷新频率

Market quotes refresh hourly. Protocol fundamentals, stablecoin supply and network statistics refresh daily. Asset profiles, developer activity and long price history rotate on a weekly and monthly cycle so that a rate-limited data allowance covers the whole universe rather than the top of it. The exact time each feed last succeeded is published on Data freshness, including failures.

对目录本身的覆盖情况

The site currently defines 124 figures across 750 assets. Nothing is displayed anywhere on the site that does not have an entry in the catalog, and nothing in the catalog is displayed without its period and source.